[Date Prev][Date Next][Thread Prev][Thread Next][Author Index][Date Index][Thread Index]

Re: :zz,osmic,f3,unix: Could a cell be an inode? [resend]

This discussion belongs on the OSMIC mailing list, so I've redirected
it there.

On Tue, Oct 06, 1998 at 02:51:13PM +1000, Gossamer wrote:
> Ted Nelson wrote:
> > Trying to set aside my dislike of HTML for a minute--
> This isn't HTML.  Don't get HTML and URLs confused - HTML is a screwy
> text markup system, URLs are a very very general way to refer to
> almost anything.  Useful.

Yes, exactly.

> Actually, something I keep forgetting to say ... I suggest that for
> OSMIC, if we seriously want to be able to access it from ZigZag
> cells, we do:
>    1.  I re-design the protocal and re-code the server.  It's not
>        very good at the moment.
>    2.  Somebody (I'm game to start!) writes up an RFC for the OSMIC
>        protocol and releases it.
>    3.  Somebody (ditto) writes up an RFC for the URLs in OSMIC style,
>        ie:
>           osmic://hostname.goes.here/2-4,5,69-87
>        where the last part is the offsets.  That means any web browser
>        that picks up and implements that RFC spec can access OSMIC
>        servers.

That would all be wonderful and very useful.  Thank you!!
BTW, I believe there are two kinds of OSMIC references:
references to versions, which return the string of pointers that
comprise that version, and pointer dereferences which return
actual content from the primedia stream.

> > >This would allow us to incorporate cells
> > >that point to files or portions of files as you suggested. 
> > How could it point to "portions of files" without
> >  embedded anchors?
> 1.  Get them off a handy OSMIC server
> 2.  Use the embedded anchors if it's already a HTML document.
> 3.  Invent something else.  If we use URLs then we can just invent
>     new ones.  Eg, off the top of my head ...  imagine this:
>        text://localhost/home/gossamer/notes#2-49
>     to retreive lines 2-49 of a text file.  (But I'm sure that's not
>     what Xanni meant!)

Close; take a look at my "text transcopyright" RFC draft.  The idea
is to specify a new fragment identifier that can be applied
to existing URIs so that you can use "file:" for local files,
plus "http:" and all other existing protocols.  Something like
"|byte[2-49,57-68],char(SJIS)[101-197]" for bytes 2-49 and 57-68,
plus Shift-JIS (a variable byte length encoding) characters 101-197.

Other kinds of selection (by line, by string match, etc.) should
also be supported.  I believe the HyTime standard already describes
a very general set of selectors that support things including multiple
dimensions (for images, 3D objects or videos) and user-defined dimensions.
Unfortunately XML has apparently only adopted the hierarchy selection
parts of this standard, not surprisingly since XML can basically only
handle documents that can be described as a hierarchy.  :-(

> > Or without resuscitating our currently-defunct
> >  TXT SRC server?
> What's that?

It's a CGI script that preprocesses a <TXT SRC> tag in HTML documents
and returns a normal HTML document with the result of evaluating all
the <TXT SRC> tags, and with all the HREFs changed to point back through
the script so that further documents retrieved will also be preprocessed.

Other ways to do this include with a web proxy (which would support
users of that proxy server) or with a web server module (which would
support the feature for pages hosted on that server).

> For that matter, what's Microcosm?

It's a hypermedia linkbase partly inspired by Xanadu; see www.multicosm.com.

Share and enjoy,
		*** Xanni ***
mailto:xanni@xxxxxxxxxx                         Andrew Pam
http://www.xanadu.com.au/                       Technical VP, Xanadu
http://www.glasswings.com.au/                   Technical Editor, Glass Wings
http://www.sericyb.com.au/sc/                   Manager, Serious Cybernetics
P.O. Box 26, East Melbourne VIC 8002 Australia  Phone +61 3 96511511